A system and method for enterprise strategy planning lifecycle closed-loop management

Abstract

The application discloses a kind of enterprise strategic planning full life cycle closed loop management and method, it is related to enterprise information management and network security technical field, by obtaining enterprise strategic planning and forming separate version after each change;Version encryption processing and obtaining hash fingerprint;Hash fingerprint is decentralized with unique version number Storage, encrypted version and key are stored in cloud server;When user accesses, network traffic characteristics are collected and input into network security analysis model trained by space limited multi-order exploration algorithm, to determine security monitoring result;If safe, dispatch target version data, and verify its integrity by the hash fingerprint of decentralized storage;After verification, data and the key encrypted using user public key are securely transmitted to user, absolute tamper-proofing and closed loop tracing of strategic planning version are realized, and the accuracy of network environment perception is improved by multi-order exploration algorithm, a highly secure data access and transmission defense line is built.

Description

Technical Field

[0001] This invention relates to the fields of enterprise information management and network security technology, specifically to a closed-loop management system and method for the entire lifecycle of enterprise strategic planning. Background Technology

[0002] Corporate strategic planning is crucial in determining a company's long-term development direction and core competitiveness. With the development of information technology, the formulation, modification, approval, and issuance of corporate strategic plans are increasingly being digitized. However, existing corporate strategic planning management systems suffer from the following technical shortcomings: Strategic plans undergo multiple adjustments throughout their lifecycle, and existing systems often only cover the currently effective version, lacking a lifecycle-wide version fixation mechanism. Furthermore, centrally stored data is at risk of unauthorized modification by internal personnel, making it difficult to clarify responsibilities. Corporate strategic plans are highly confidential information, and existing systems lack in-depth security testing of the access environment (such as network traffic characteristics) when users access them, making them vulnerable to threats such as man-in-the-middle attacks and malicious sniffing. Traditional network security analysis models suffer from slow convergence speeds and a tendency to get trapped in local optima when facing complex and ever-changing network attack traffic, resulting in low security monitoring accuracy and a high likelihood of false positives or false negatives. Summary of the Invention

[0003] The purpose of this invention is to provide a closed-loop management system and method for the entire lifecycle of enterprise strategic planning, in order to solve the problems of easy tampering of strategic planning versions, insecure access environment, and low accuracy of security detection models in the prior art.

[0004] The first aspect of this invention provides a closed-loop management system for the entire lifecycle of enterprise strategic planning, comprising: The Enterprise Strategic Planning Management module is used to acquire enterprise strategic plans generated by enterprise users through human-computer interaction, and to generate a separate version of the enterprise strategic plan each time the enterprise strategic plan changes. The data encryption management module is used to encrypt the enterprise strategic planning version through the cloud server to obtain the encrypted enterprise strategic planning version, and at the same time obtain the hash fingerprint of the encrypted enterprise strategic planning version. The blockchain storage module is used for decentralized storage of the hash fingerprint of the encrypted version of the enterprise strategic plan and its corresponding unique version number. The data storage module is used to store the encrypted version of the enterprise strategic plan, the encryption key, and the corresponding unique version number on the cloud server. The cloud service security management module is used to collect the network access environment characteristics of enterprise users when they access any version of the enterprise strategic plan, and to analyze the network access environment characteristics to determine the cloud service security monitoring results. The cloud service security management module is also used to schedule the corresponding encrypted version of the enterprise strategic planning when the cloud service security monitoring result is secure, to obtain the target version of the enterprise strategic planning data; to verify the target version of the enterprise strategic planning data by using the hash fingerprint of the encrypted version of the enterprise strategic planning and the corresponding unique version number stored in a decentralized manner; and to securely transmit the target version of the enterprise strategic planning data and the encryption key to the enterprise user if the verification is successful.

[0005] A second aspect of this invention provides a closed-loop management method for the entire lifecycle of enterprise strategic planning, comprising: Acquire enterprise strategic plans generated by enterprise users through human-computer interaction, and generate a separate version of the enterprise strategic plan each time the enterprise strategic plan changes; The enterprise strategic planning version is encrypted using a cloud server to obtain an encrypted version of the enterprise strategic planning, and the hash fingerprint of the encrypted version of the enterprise strategic planning is also obtained. The hash fingerprint and corresponding unique version number of the encrypted enterprise strategic plan version are stored in a decentralized manner. The encrypted version of the enterprise strategic plan, the encryption key, and the corresponding unique version number are stored on a cloud server. When an enterprise user accesses any version of the enterprise strategic plan, the network access environment characteristics of the enterprise user are collected, and the network access environment characteristics are analyzed to determine the cloud service security monitoring results. If the cloud service security monitoring result is secure, then the corresponding encrypted version of the enterprise strategic plan will be scheduled to obtain the target version of the enterprise strategic plan data. The target version of the enterprise strategic planning data is verified by using the hash fingerprint of the encrypted version of the enterprise strategic planning data and the corresponding unique version number stored in a decentralized manner. If the verification is successful, the target version of the enterprise strategic planning data and the encryption key are securely transmitted to the enterprise user.

[0006] Furthermore, when an enterprise user accesses any version of the enterprise strategic plan, the network access environment characteristics of the enterprise user are collected, and the network access environment characteristics are analyzed to determine the cloud service security monitoring results, including: When an enterprise user accesses any version of the enterprise strategic plan, the network traffic characteristics of the enterprise user during the access process are collected to obtain the network access environment characteristics of the enterprise user. The network access environment characteristics of the enterprise users are input into the network security analysis model trained by the intelligent optimization algorithm to determine the cloud service security monitoring results output by the network security analysis model.

[0007] Furthermore, the training method for a network security analysis model trained by a space-constrained multi-order exploration algorithm includes: A network security analysis model is constructed, and the model parameters of the network security analysis model are encoded to obtain particle swarm optimization. For any particle in the particle swarm, obtain the fitness of the particle and determine the particle with the highest fitness as the optimal particle. A gradient neighborhood exploration mechanism is used to quickly explore the neighborhood information of particles, resulting in particles after rapid exploration of neighborhood information. Based on the optimal particle, a local solution space exploration is performed on the particle after rapid exploration of neighborhood information using an exploration space-limited exploration mechanism to obtain the particle after local solution space exploration. A multi-particle information adaptive mixing mechanism is used to supplement the local solution space exploration of the particles after the local solution space exploration, so as to obtain the particles after the local solution space supplementation exploration. A solution space resampling mechanism is used to explore the global solution space of the particles after supplementing the exploration of the local solution space, so as to obtain the particles after exploring the global solution space. Repeatedly execute gradient neighborhood order exploration mechanism, exploration space-limited exploration mechanism, multi-particle information adaptive mixing mechanism, and solution space resampling mechanism; Until the maximum number of training iterations is reached, the optimal particle is redefined, and the model parameters in the optimal particle are used as the target model parameters of the network security analysis model, thus obtaining the network security analysis model trained by the spatially constrained multi-order exploration algorithm.

[0008] Furthermore, a network security analysis model is constructed, and the model parameters of the network security analysis model are encoded to obtain particle swarm optimization, including: Convolutional neural networks, long short-term memory networks, or a combination of both are used to construct network security analysis models. The model parameters of the network security analysis model are randomly initialized between the upper limit and the lower limit of the model parameters and encoded into a vector to obtain the initial particles; Based on the initial particle, several other particles are obtained as follows: in, Indicates the first i The first particle d dimensional model parameters, and i When it is 1, Represents the initial particle. Indicates the first i +1 particle's first d The model parameters are d=1,2,...,D, where D represents the total dimension of the model parameters. Represents the modulo function. Represents pi (π). This represents the mapping factor between (0,1). This represents the piecewise coefficient between (0, 0.5). This represents the first random number between (0, 1).

[0009] Furthermore, a gradient neighborhood exploration mechanism is employed to rapidly explore the neighborhood information of particles, resulting in particles after rapid neighborhood information exploration, including: The neighborhood gradient information of the particle is obtained as follows: in, In the t-th training process, the first... n One particle, This represents the neighborhood gradient information of the nth particle during the t-th training process. This represents a random particle where each element is randomly set to either 0.1 or -0.1. express Corresponding fitness express fitness Represents a second random number between (0, 1); Based on the neighborhood gradient information, a rapid neighborhood information exploration is performed on the particles to obtain the particles after rapid neighborhood information exploration: in, Indicates the first n Particles after rapid exploration of neighborhood information, This represents the neighborhood information exploration speed control factor. This represents a parameter vector where all elements are 1. Represents a symbolic function. This represents the base speed and is set to 0.1. This represents the parameter that controls the influence of the factor, and is set to a random number between (0.1, 0.45). Represents pi (π). Indicates the maximum number of training iterations. This represents the total number of particles.

[0010] Furthermore, based on the optimal particle, a local solution space exploration is performed on the particle after rapid exploration of neighborhood information using an exploration space-constrained exploration mechanism, resulting in the particle after local solution space exploration, including: Based on the aforementioned optimal particle, the exploration space is limited to: in, Indicates the lower bound of the hyperparameter. Indicates the lower bound of the hyperparameter. Indicates the lower limit of the exploration space. Indicates the upper limit of the exploration space. This indicates taking the maximum value. This indicates taking the minimum value. Let d represent the d-th dimension of the model parameters for the optimal particle. =t / T, Indicates the search space constraint factor; Based on the aforementioned exploration space, the particles after rapid exploration of neighborhood information are subjected to local solution space exploration, resulting in the following particles after local solution space exploration: in, Indicates the first t During the training process, the first k Particles after rapid exploration of neighborhood information, Indicates the first k The particle after exploring a local solution space, where e represents the natural constant. Indicates the exploration path control factor. This represents the exploration direction control factor between (-1, 1). This represents a constant term, and is set to 5.

[0011] Furthermore, a multi-particle information adaptive mixing mechanism is employed to supplement the local solution space exploration of the particles after the local solution space exploration, resulting in particles after the supplementary exploration of the local solution space, including: For the particles after exploring the local solution space, three other particles after exploring the local solution space are randomly matched to obtain the first matching particle, the second matching particle, and the third matching particle. Based on the first matching particle, the second matching particle, and the third matching particle, the first information mixing factor and the second information mixing factor are obtained as follows: in, Indicates the first m The first information mixing factor corresponding to the d-th dimension model parameters of the particle after exploring a local solution space. Indicates the first m The second information mixing factor corresponding to the d-th dimension model parameters of the particle after exploring a local solution space. This represents the d-th dimension of the model parameters for the first matched particle. This represents the d-th dimension of the model parameters for the second matched particle. This represents the d-th dimension of the model parameters for the third matched particle. This indicates the decimal part and is set to 0.01 or 0.001. This represents the first parameter for evaluating training effectiveness. This represents the second parameter for evaluating training effectiveness. This represents the third parameter for evaluating training effectiveness. The training effectiveness evaluation parameter = | / +0.4t / T; Let represent the average fitness of the particle after exploring all local solution spaces during the t-th training iteration. This represents the average fitness of the matched particles during the training process; Based on the first information mixing factor and the second information mixing factor, the particles after the local solution space exploration are further explored in the local solution space to obtain the following particles after the local solution space further exploration: in, Indicates the first t During the training process, the first m The d-th dimension model parameters of the particle after exploring a local solution space. Indicates the first m The local solution space is used to supplement the d-th dimension model parameters of the particle after exploration. Represents the natural constant.

[0012] Furthermore, a solution space resampling mechanism is used to perform a global solution space exploration on the particles after the local solution space is supplemented, resulting in the following particles after the global solution space exploration: in, Indicates the first t During the training process, the first i A local solution space is used to supplement the particles after exploration. Indicates the first i Particles after exploring a global solution space. Represents the global exploration coefficient. H represents the resampling index, and H represents the resampling order. This represents the global exploration control factor, and is set to 0.45. Represents the optimal particle. Represented as Other randomly matched local solution spaces supplement the particles explored after the initial exploration.

[0013] Furthermore, the target version of the enterprise strategic planning data is verified using the hash fingerprint of the encrypted version and its corresponding unique version number stored in a decentralized manner. Upon successful verification, the target version of the enterprise strategic planning data and the encryption key are securely transmitted to the enterprise user, including: Based on the unique version number corresponding to the target version of the enterprise strategic planning data, a unique version number is obtained by matching the unique version number stored in a decentralized manner. The target hash fingerprint is obtained by finding the hash fingerprint of the encrypted enterprise strategic planning version stored in a decentralized manner using the target unique version number. Obtain the real-time hash fingerprint corresponding to the target version of the enterprise strategic planning data, and verify the real-time hash fingerprint with the target hash fingerprint to obtain the verification result; the verification result includes verification pass or verification failure. If the verification result is a verification failure, the current process will end and an error message will be sent to the enterprise user. If the verification result is successful, the encryption key is encrypted using the enterprise user's public key to obtain the key information; The target version of the enterprise strategic planning data and key information will be transmitted to the enterprise user.

[0014] The beneficial effects of this invention are as follows: This invention provides a closed-loop management method for the entire lifecycle of enterprise strategic planning. It involves acquiring the enterprise strategic plan and creating a separate version after each change; encrypting the version and obtaining a hash fingerprint; storing the hash fingerprint and unique version number in a decentralized manner; and storing the encrypted version and key on a cloud server. When a user accesses the platform, network traffic characteristics are collected and input into a network security analysis model trained by a spatially limited multi-stage exploration algorithm to determine the security monitoring results. If secure, the target version data is scheduled, and its integrity is verified using the decentralized hash fingerprint. After successful verification, the data and the key encrypted with the user's public key are securely transmitted to the user. This invention achieves absolute tamper-proof and closed-loop traceability of strategic planning versions and improves the accuracy of network environment awareness through a multi-stage exploration algorithm, constructing a highly secure data access and transmission defense line. Attached Figure Description

[0015] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with the invention and, together with the description, serve to explain the principles of the invention.

[0016] Figure 1A flowchart of a closed-loop management method for the entire lifecycle of enterprise strategic planning provided by this invention; Figure 2 A structural diagram of a closed-loop management system for the entire lifecycle of enterprise strategic planning provided by this invention; Among them, 201-Enterprise Strategic Planning Management Module, 202-Data Encryption Management Module, 203-Blockchain Storage Module, 204-Data Storage Module, 205-Cloud Service Security Management Module, and 206-Cloud Service Security Management Module; The accompanying drawings have illustrated specific embodiments of the invention, which will be described in more detail below. These drawings and descriptions are not intended to limit the scope of the invention in any way, but rather to illustrate the concept of the invention to those skilled in the art through reference to particular embodiments. Detailed Implementation

[0017] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numerals in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present invention. Rather, they are merely examples of apparatuses and methods consistent with some aspects of the invention as detailed in the appended claims.

[0018] The embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0019] like Figure 1 As shown, this embodiment of the invention provides a closed-loop management method for the entire lifecycle of enterprise strategic planning, including: S101. Obtain the enterprise strategic plan generated by enterprise users through human-computer interaction, and form a separate version of the enterprise strategic plan after each change in the enterprise strategic plan; S102. The enterprise strategic planning version is encrypted using a cloud server to obtain an encrypted version of the enterprise strategic planning, and the hash fingerprint of the encrypted version of the enterprise strategic planning is obtained at the same time. S103. Decentralize the storage of the hash fingerprint and corresponding unique version number of the encrypted enterprise strategic planning version. S104. Store the encrypted version of the enterprise strategic plan, the encryption key, and the corresponding unique version number on the cloud server. For example, the system obtains enterprise strategic plans generated by enterprise users through human-computer interaction, and forms a separate version of the enterprise strategic plan after each change; it obtains hash fingerprints through cloud server encryption; the hash fingerprints and version numbers are uploaded to the blockchain; and the ciphertext, key, and version number are stored in the cloud.

[0020] S105. When an enterprise user accesses any version of the enterprise strategic plan, the network access environment characteristics of the enterprise user are collected, and the network access environment characteristics are analyzed to determine the cloud service security monitoring results. When an enterprise user accesses any version of the enterprise strategic plan, the network traffic characteristics of the enterprise user during the access process (such as packet length distribution, flow duration, TCP flag ratio, etc.) are collected to obtain network access environment characteristics; these characteristics are input into a network security analysis model trained by a spatially limited multi-level exploration algorithm to output cloud service security monitoring results (secure / insecure).

[0021] S106. If the cloud service security monitoring result is secure, then schedule the corresponding encrypted version of the enterprise strategic plan to obtain the target version of the enterprise strategic plan data. If the monitoring result indicates insecurity, access is denied; if secure, the corresponding encrypted ciphertext (target version of enterprise strategic planning data) is retrieved from the cloud. Integrity verification is then performed: the target version number is matched against the unique target version number on the blockchain to obtain the target hash fingerprint; a real-time hash fingerprint is calculated and compared with the target hash fingerprint. If verification fails, the process ends and an exception is reported (indicating that the data may have been tampered with); if verification succeeds, the encryption key is encrypted using the public key pre-provided by the enterprise user to obtain the key information, and the ciphertext and key information are jointly sent to the enterprise user's client. The user uses their private key to decrypt the key and then decrypts it to obtain the plaintext strategic planning data.

[0022] S107. Verify the target version of the enterprise strategic planning data by using the hash fingerprint of the encrypted version of the enterprise strategic planning data and the corresponding unique version number stored in a decentralized manner, and if the verification is successful, securely transmit the target version of the enterprise strategic planning data and the encryption key to the enterprise user.

[0023] The cloud-based blockchain's dual-layer storage architecture stores the complete strategic plan in encrypted form on a cloud server, while storing its hash fingerprint on the blockchain. Due to the immutable and decentralized nature of blockchain, any unauthorized modification to the cloud-based strategic plan data will be immediately detected during hash comparison. Furthermore, each change generates a unique version number, enabling closed-loop traceability throughout the entire lifecycle of the strategy, from its proposal and iteration to its eventual abandonment.

[0024] This invention does not simply encrypt and transmit data, but incorporates in-depth analysis of the user's network access environment characteristics (traffic characteristics) before data scheduling, blocking data transmission under malicious eavesdropping conditions at the source; during data transmission, asymmetric encryption is used to protect the key, achieving secure isolation transmission of the key and ciphertext, greatly improving the security of core data.

[0025] In some possible implementations, when an enterprise user accesses any version of the enterprise strategic plan, the network access environment characteristics of the enterprise user are collected, and the network access environment characteristics are analyzed to determine the cloud service security monitoring results, including: When an enterprise user accesses any version of the enterprise strategic plan, the network traffic characteristics of the enterprise user during the access process are collected to obtain the network access environment characteristics of the enterprise user. The network access environment characteristics of the enterprise users are input into the network security analysis model trained by the intelligent optimization algorithm to determine the cloud service security monitoring results output by the network security analysis model.

[0026] In some possible implementations, the training method for a network security analysis model trained by a space-constrained multi-order exploration algorithm includes: A network security analysis model is constructed, and the model parameters of the network security analysis model are encoded to obtain particle swarm optimization. For any particle in the particle swarm, obtain the fitness of the particle and determine the particle with the highest fitness as the optimal particle; for example, the loss function value of the particle can be obtained by using a dataset such as KDD 99 flow, and the fitness can be obtained by adding the loss function value to a preset constant (such as 1) and taking the reciprocal. A gradient neighborhood exploration mechanism is used to quickly explore the neighborhood information of particles, resulting in particles after rapid exploration of neighborhood information. Based on the optimal particle, a local solution space exploration is performed on the particle after rapid exploration of neighborhood information using an exploration space-limited exploration mechanism to obtain the particle after local solution space exploration. A multi-particle information adaptive mixing mechanism is used to supplement the local solution space exploration of the particles after the local solution space exploration, so as to obtain the particles after the local solution space supplementation exploration. A solution space resampling mechanism is used to explore the global solution space of the particles after supplementing the exploration of the local solution space, so as to obtain the particles after exploring the global solution space. Repeatedly execute gradient neighborhood order exploration mechanism, exploration space-limited exploration mechanism, multi-particle information adaptive mixing mechanism, and solution space resampling mechanism; The process continues until the maximum number of training iterations is reached. Then, the optimal particle is redefined, and the model parameters from this optimal particle are used as the target model parameters for the network security analysis model. This results in a network security analysis model trained using a space-constrained multi-stage exploration algorithm. For example, in the optimal training iteration, the particle with the highest fitness after exploring the global solution space is selected as the optimal particle.

[0027] The spatially constrained multi-order exploration algorithm addresses the problem of network security analysis models easily getting trapped in local optima: traditional algorithms such as particle swarm optimization often face a contradiction between convergence and optimization when training security analysis models. This invention innovatively designs four step-by-step mechanisms: gradient neighborhood exploration, spatially constrained exploration, adaptive mixing of multi-particle information, and solution space resampling. Gradient neighborhood exploration utilizes a sign function and an adaptive velocity control factor to achieve efficient, targeted, and rapid exploration of neighborhood information. Exploration space-bounded exploration dynamically shrinks the search boundary centered on the current optimal particle, significantly improving local optimization accuracy. Adaptive multi-particle information mixing dynamically adjusts the information mixing factor by introducing training effect evaluation parameters, achieving deep cross-fertilization of excellent features between particles. Solution space resampling introduces a high-order resampling index and control factor, endowing particles with the ability to explore globally and escape local extrema. This algorithm enables the trained network security analysis model to achieve higher accuracy (Acc) and lower false positive rate when identifying complex network attack traffic, providing an extremely reliable security gateway for cloud service access.

[0028] In some possible implementations, a network security analysis model is constructed, and the model parameters of the network security analysis model are encoded to obtain a particle swarm optimization, including: Convolutional neural networks, long short-term memory networks, or a combination of both are used to construct network security analysis models. Preferably, a network security analysis model can be constructed by combining a convolutional neural network (CNN) and a long short-term memory network (LSTM), with CNN extracting spatial traffic features and LSTM extracting temporal traffic features.

[0029] The model parameters of the network security analysis model are randomly initialized between the upper limit and the lower limit of the model parameters and encoded into a vector to obtain the initial particles; Based on the initial particle, several other particles are obtained as follows: in, Indicates the first i The first particle d dimensional model parameters, and i When it is 1, Represents the initial particle. Indicates the first i +1 particle's first d The model parameters are d=1,2,...,D, where D represents the total dimension of the model parameters. Represents the modulo function. Represents pi (π). This represents the mapping factor between (0,1). This represents the piecewise coefficient between (0, 0.5). This represents the first random number between (0, 1).

[0030] This segmentation mechanism ensures that the initial particles are distributed extremely evenly in the solution space, avoiding the aggregation phenomenon caused by traditional random initialization.

[0031] In some possible implementations, a gradient neighborhood exploration mechanism is used to quickly explore the neighborhood information of the particle, resulting in a particle after the rapid exploration of neighborhood information, including: The neighborhood gradient information of the particle is obtained as follows: in, In the t-th training process, the first... n One particle, This represents the neighborhood gradient information of the nth particle during the t-th training process. This represents a random particle where each element is randomly set to either 0.1 or -0.1. express Corresponding fitness express fitness This represents the second random number between (0,1); the formula simulates the numerical gradient through a small perturbation, avoiding the huge amount of computation required to calculate the analytical gradient of a complex neural network.

[0032] Based on the neighborhood gradient information, a rapid neighborhood information exploration is performed on the particles to obtain the particles after rapid neighborhood information exploration: in, Indicates the first n Particles after rapid exploration of neighborhood information, This represents the neighborhood information exploration speed control factor. This represents a parameter vector where all elements are 1. Represents a symbolic function. This represents the base speed and is set to 0.1. This represents the parameter that controls the influence of the factor, and is set to a random number between (0.1, 0.45). Represents pi (π). Indicates the maximum number of training iterations. This represents the total number of particles.

[0033] Gradient neighborhood exploration utilizes a sign function and an adaptive velocity control factor to achieve efficient, targeted, and rapid exploration of neighborhood information. The design allows the algorithm to have a large step size for rapid exploration in the early stages of training (when t is small). As training progresses, the step size adaptively decreases, and the sign function sign ensures that the exploration direction always follows the direction of increasing fitness.

[0034] In some possible implementations, based on the optimal particle, a local solution space exploration is performed on the particle after rapid exploration of neighborhood information using an exploration space-limited exploration mechanism, resulting in the particle after local solution space exploration, including: Based on the aforementioned optimal particle, the exploration space is limited to: in, Indicates the lower bound of the hyperparameter. Indicates the lower bound of the hyperparameter. Indicates the lower limit of the exploration space. Indicates the upper limit of the exploration space. This indicates taking the maximum value. This indicates taking the minimum value. Let d represent the d-th dimension of the model parameters for the optimal particle. =t / T, This represents the search space constraint factor; as the number of training iterations increases, As the exploration space gradually shrinks from 0 to near 1, it gradually shrinks to the vicinity of the optimal particle. Based on the aforementioned exploration space, the particles after rapid exploration of neighborhood information are subjected to local solution space exploration, resulting in the following particles after local solution space exploration: in, Indicates the first t During the training process, the first k Particles after rapid exploration of neighborhood information, Indicates the first k The particle after exploring a local solution space, where e represents the natural constant. Indicates the exploration path control factor. This represents the exploration direction control factor between (-1, 1). This represents a constant term, and is set to 5.

[0035] The exploration space is limited by dynamically shrinking the search boundary with the current best particle as the center, which greatly improves the accuracy of local optimization. By introducing the cosine function and the exploration path control factor, the particle generates high-frequency oscillations within the limited solution space, which can strip away redundant parameters around the local extrema with great precision.

[0036] In some possible implementations, a multi-particle information adaptive hybridization mechanism is used to perform local solution space supplementation exploration on the particles after local solution space exploration, resulting in particles after local solution space supplementation exploration, including: For the particles after exploring the local solution space, three other particles after exploring the local solution space are randomly matched to obtain the first matching particle, the second matching particle, and the third matching particle. Based on the first matching particle, the second matching particle, and the third matching particle, the first information mixing factor and the second information mixing factor are obtained as follows: in, Indicates the first m The first information mixing factor corresponding to the d-th dimension model parameters of the particle after exploring a local solution space. Indicates the first m The second information mixing factor corresponding to the d-th dimension model parameters of the particle after exploring a local solution space. This represents the d-th dimension of the model parameters for the first matched particle. This represents the d-th dimension of the model parameters for the second matched particle. This represents the d-th dimension of the model parameters for the third matched particle. This indicates the decimal part and is set to 0.01 or 0.001. This represents the first parameter for evaluating training effectiveness. This represents the second parameter for evaluating training effectiveness. This represents the third parameter for evaluating training effectiveness. The training effectiveness evaluation parameter = | / +0.4t / T; Let represent the average fitness of the particle after exploring all local solution spaces during the t-th training iteration. This represents the average fitness of the matched particles during the training process; Based on the first information mixing factor and the second information mixing factor, the particles after the local solution space exploration are further explored in the local solution space to obtain the following particles after the local solution space further exploration: in, Indicates the first t During the training process, the first m The d-th dimension model parameters of the particle after exploring a local solution space. Indicates the first m The local solution space is used to supplement the d-th dimension model parameters of the particle after exploration. Represents the natural constant.

[0037] Multi-particle information adaptive mixing dynamically adjusts the information mixing factor by introducing training effect evaluation parameters, achieving deep cross-fertilization of excellent features among particles. This nonlinear exponential update strategy breaks the limitations of traditional linear cross-fertilization, enabling the inheritance of excellent genes to exhibit exponential explosive power, effectively ensuring particle diversity and improving the exploration rate in unknown regions of the solution space.

[0038] In some possible implementations, a solution space resampling mechanism is used to perform a global solution space exploration on the particles after the local solution space has been supplemented, resulting in the following particles after the global solution space exploration: in, Indicates the first t During the training process, the first i A local solution space is used to supplement the particles after exploration. Indicates the first i Particles after exploring a global solution space. Represents the global exploration coefficient. H represents the resampling index, and H represents the resampling order. This represents the global exploration control factor, and is set to 0.45. Represents the optimal particle. Represented as Other randomly matched local solution spaces supplement the particles explored after the initial exploration.

[0039] Solution space resampling introduces higher-order resampling indices and control factors, giving particles the ability to explore globally and escape local optima. The effect of this mechanism itself naturally decays and will not destroy the excellent solutions already found. When encountering local optima in the middle stage, the introduction of higher-order terms will produce parameter space jumps (similar to quantum leaps). Combined with the differential guidance of optimal and random particles, this can pull particles out of local traps with a very high probability and into new, promising solution spaces.

[0040] In some possible implementations, the target version of the enterprise strategic planning data is verified using the hash fingerprint of the encrypted version of the enterprise strategic planning data and its corresponding unique version number stored in a decentralized manner. Upon successful verification, the target version of the enterprise strategic planning data and the encryption key are securely transmitted to the enterprise user, including: Based on the unique version number corresponding to the target version of the enterprise strategic planning data, a unique version number is obtained by matching the unique version number stored in a decentralized manner. The target hash fingerprint is obtained by finding the hash fingerprint of the encrypted enterprise strategic planning version stored in a decentralized manner using the target unique version number. Obtain the real-time hash fingerprint corresponding to the target version of the enterprise strategic planning data, and verify the real-time hash fingerprint with the target hash fingerprint to obtain the verification result; the verification result includes verification pass or verification failure. If the verification result is a verification failure, the current process will end and an error message will be sent to the enterprise user. If the verification result is successful, the encryption key is encrypted using the enterprise user's public key to obtain the key information; The target version of the enterprise strategic planning data and key information will be transmitted to the enterprise user.

[0041] like Figure 2 As shown, the present invention provides a closed-loop management system for the entire lifecycle of enterprise strategic planning, comprising: The Enterprise Strategic Planning Management Module 201 is used to acquire enterprise strategic plans generated by enterprise users through human-computer interaction, and to generate a separate version of the enterprise strategic plan each time the enterprise strategic plan changes. The data encryption management module 202 is used to encrypt the enterprise strategic planning version through the cloud server to obtain the encrypted enterprise strategic planning version, and at the same time obtain the hash fingerprint of the encrypted enterprise strategic planning version. Blockchain storage module 203 is used for decentralized storage of the hash fingerprint of the encrypted version of the enterprise strategic plan and its corresponding unique version number. Data storage module 204 is used to store the encrypted version of the enterprise strategic plan, the encryption key, and the corresponding unique version number on a cloud server; The cloud service security management module 205 is used to collect the network access environment characteristics of enterprise users when they access any version of the enterprise strategic plan, and to analyze the network access environment characteristics to determine the cloud service security monitoring results. The cloud service security management module 206 is also used to, when the cloud service security monitoring result is secure, schedule the corresponding encrypted version of the enterprise strategic planning to obtain the target version of the enterprise strategic planning data; verify the target version of the enterprise strategic planning data by using the hash fingerprint of the encrypted version of the enterprise strategic planning in decentralized storage and the corresponding unique version number; and, if the verification is successful, securely transmit the target version of the enterprise strategic planning data and the encryption key to the enterprise user.

[0042] The enterprise strategic planning full life cycle closed-loop management system provided in this embodiment of the invention can execute the technical solutions described in any of the above method embodiments. Its principle and beneficial effects are similar, and will not be repeated here.

[0043] Embodiments of the present invention are described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processing unit of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processing unit of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0044] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0045] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0046] Other embodiments of the invention will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This invention is intended to cover any variations, uses, or adaptations of the invention that follow the general principles of the invention and include common knowledge or customary techniques in the art not disclosed herein. The specification and examples are to be considered exemplary only, and the true scope and spirit of the invention are indicated by the claims.

[0047] It should be understood that the present invention is not limited to the precise structure described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of the invention is limited only by the appended claims.

Claims

1. A closed-loop management method for the entire lifecycle of enterprise strategic planning, characterized in that, include: Acquire enterprise strategic plans generated by enterprise users through human-computer interaction, and create a separate version of the enterprise strategic plan each time the enterprise strategic plan changes; The enterprise strategic planning version is encrypted using a cloud server to obtain an encrypted version of the enterprise strategic planning, and the hash fingerprint of the encrypted version of the enterprise strategic planning is also obtained at the same time. The hash fingerprint and corresponding unique version number of the encrypted enterprise strategic plan version are stored in a decentralized manner. The encrypted version of the enterprise strategic plan, the encryption key, and the corresponding unique version number are stored on a cloud server; When an enterprise user accesses any version of the enterprise strategic plan, the network access environment characteristics of the enterprise user are collected, and the network access environment characteristics of the enterprise user are input into a network security analysis model trained by a spatially limited multi-level exploration algorithm to determine the cloud service security monitoring results output by the network security analysis model. If the cloud service security monitoring result is secure, then the corresponding encrypted version of the enterprise strategic plan will be scheduled to obtain the target version of the enterprise strategic plan data. The target version of the enterprise strategic planning data is verified by the hash fingerprint of the encrypted version of the enterprise strategic planning and the corresponding unique version number stored in a decentralized manner. If the verification is successful, the target version of the enterprise strategic planning data and the encryption key are securely transmitted to the enterprise user. Constructing the network security analysis model trained by the spatially constrained multi-order exploration algorithm includes: A network security analysis model is constructed, and the model parameters of the network security analysis model are encoded to obtain particle swarm optimization. For any particle in the particle swarm, obtain the fitness of the particle and determine the particle with the highest fitness as the optimal particle. A gradient neighborhood exploration mechanism is used to quickly explore the neighborhood information of particles, resulting in particles after rapid exploration of neighborhood information. Based on the optimal particle, a local solution space exploration is performed on the particle after rapid exploration of neighborhood information using an exploration space-limited exploration mechanism to obtain the particle after local solution space exploration; A multi-particle information adaptive mixing mechanism is used to supplement the local solution space exploration of the particles after the local solution space exploration, so as to obtain the particles after the local solution space supplementation exploration. A solution space resampling mechanism is used to explore the global solution space of the particles after supplementing the exploration of the local solution space, so as to obtain the particles after exploring the global solution space. Repeatedly execute gradient neighborhood order exploration mechanism, exploration space-limited exploration mechanism, multi-particle information adaptive mixing mechanism, and solution space resampling mechanism; Until the maximum number of training iterations is reached, the optimal particle is redefined, and the model parameters in the optimal particle are used as the target model parameters of the network security analysis model, thus obtaining the network security analysis model trained by the spatially constrained multi-order exploration algorithm.

2. The closed-loop management method for the entire lifecycle of enterprise strategic planning according to claim 1, characterized in that, Construct a network security analysis model, encode the model parameters of the network security analysis model, and obtain particle swarm optimization, including: Convolutional neural networks, long short-term memory networks, or a combination of both are used to construct network security analysis models. The model parameters of the network security analysis model are randomly initialized between the upper limit and the lower limit of the model parameters and encoded into a vector to obtain the initial particles; Based on the initial particle, several other particles are obtained as follows: in, Indicates the first i The first particle d dimensional model parameters, and i When it is 1, Represents the initial particle. Indicates the first i +1 particle's first d The model parameters are d=1,2,...,D, where D represents the total dimension of the model parameters. Represents the modulo function. Represents pi (π). This represents the mapping factor between (0,1). This represents the piecewise coefficient between (0, 0.5). This represents the first random number between (0,1).

3. The closed-loop management method for the entire lifecycle of enterprise strategic planning according to claim 1, characterized in that, A gradient neighborhood exploration mechanism is used to quickly explore the neighborhood information of particles, resulting in particles after rapid neighborhood information exploration, including: The neighborhood gradient information of the particle is obtained as follows: in, In the t-th training process, the first... n One particle, This represents the neighborhood gradient information of the nth particle during the t-th training process. This represents a random particle where each element is randomly set to either 0.1 or -0.

1. express Corresponding fitness express Adaptability, Represents a second random number between (0, 1); Based on the neighborhood gradient information, a rapid neighborhood information exploration is performed on the particles to obtain the particles after rapid neighborhood information exploration: in, Indicates the first n Particles after rapid exploration of neighborhood information, This represents the control factor for the speed of neighborhood information exploration. This represents a parameter vector where all elements are 1. Represents a symbolic function. This represents the base speed and is set to 0.

1. This represents the parameter that controls the influence of the factor, and is set to a random number between (0.1, 0.45). Represents pi (π). Indicates the maximum number of training iterations. This represents the total number of particles.

4. The closed-loop management method for the entire lifecycle of enterprise strategic planning according to claim 3, characterized in that, Based on the optimal particle, a local solution space exploration is performed on the particle after rapid exploration of neighborhood information using an exploration space-constrained exploration mechanism, resulting in the particle after local solution space exploration, including: Based on the aforementioned optimal particle, the exploration space is limited to: in, Indicates the lower bound of the hyperparameter. This indicates the upper limit of the hyperparameters. Indicates the lower limit of the exploration space. Indicates the upper limit of the exploration space. This indicates taking the maximum value. This indicates taking the minimum value. Let d represent the d-th dimension of the model parameters for the optimal particle. =t / T, Indicates the search space constraint factor; Based on the aforementioned exploration space, the particles after rapid exploration of neighborhood information are subjected to local solution space exploration, resulting in the following particles after local solution space exploration: in, Indicates the first t During the training process, the first k Particles after rapid exploration of neighborhood information, Indicates the first k The particle after exploring a local solution space, where e represents the natural constant. Indicates the exploration path control factor. This represents the exploration direction control factor between (-1, 1). This represents a constant term, and is set to 5.

5. The closed-loop management method for the entire lifecycle of enterprise strategic planning according to claim 4, characterized in that, A multi-particle information adaptive mixing mechanism is used to supplement the local solution space exploration of the particles after the local solution space exploration, resulting in particles after the supplementary exploration of the local solution space, including: For the particles after exploring the local solution space, three other particles after exploring the local solution space are randomly matched to obtain the first matching particle, the second matching particle, and the third matching particle. Based on the first matching particle, the second matching particle, and the third matching particle, the first information mixing factor and the second information mixing factor are obtained as follows: in, Indicates the first m The first information mixing factor corresponding to the d-th dimension model parameters of the particle after exploring a local solution space. Indicates the first m The second information mixing factor corresponding to the d-th dimension model parameters of the particle after exploring a local solution space. This represents the d-th dimension of the model parameters for the first matched particle. This represents the d-th dimension of the model parameters for the second matched particle. This represents the d-th dimension of the model parameters for the third matched particle. This indicates the decimal part and is set to 0.01 or 0.

001. This represents the first parameter for evaluating training effectiveness. This represents the second parameter for evaluating training effectiveness. This represents the third parameter for evaluating training effectiveness. The training effectiveness evaluation parameter = | / +0.4t / T; Let represent the average fitness of the particle after exploring all local solution spaces during the t-th training iteration. This represents the average fitness of the matched particles during the training process; Based on the first information mixing factor and the second information mixing factor, the particles after the local solution space exploration are further explored in the local solution space to obtain the following particles after the local solution space further exploration: in, Indicates the first t During the training process, the first m The d-th dimension model parameters of the particle after exploring a local solution space. Indicates the first m The local solution space is used to supplement the d-th dimension model parameters of the particle after exploration. Represents the natural constant.

6. The closed-loop management method for the entire lifecycle of enterprise strategic planning according to claim 5, characterized in that, A solution space resampling mechanism is used to explore the global solution space of the particles after supplementing the exploration of the local solution space, resulting in the following particles after global solution space exploration: in, Indicates the first t During the training process, the first i A local solution space is used to supplement the particles after exploration. Indicates the first i Particles after exploring a global solution space. Represents the global exploration coefficient. H represents the resampling index, and H represents the resampling order. This represents the global exploration control factor, and is set to 0.

45. Represents the optimal particle. Represented as Other randomly matched local solution spaces supplement the particles explored after the initial exploration.

7. The closed-loop management method for the entire lifecycle of enterprise strategic planning according to claim 1, characterized in that, The target version of the enterprise strategic planning data is verified using the hash fingerprint of the encrypted version and its corresponding unique version number stored in a decentralized manner. Upon successful verification, the target version of the enterprise strategic planning data and the encryption key are securely transmitted to the enterprise user, including: Based on the unique version number corresponding to the target version of the enterprise strategic planning data, a unique version number is obtained by matching the unique version number stored in a decentralized manner. The target hash fingerprint is obtained by finding the hash fingerprint of the encrypted enterprise strategic planning version stored in a decentralized manner using the target unique version number. Obtain the real-time hash fingerprint corresponding to the target version of the enterprise strategic planning data, and verify the real-time hash fingerprint with the target hash fingerprint to obtain the verification result; the verification result includes verification pass or verification failure. If the verification result is a verification failure, the current process will end and an error message will be sent to the enterprise user. If the verification result is successful, the encryption key is encrypted using the enterprise user's public key to obtain the key information; The target version of the enterprise strategic planning data and key information will be transmitted to the enterprise user.

Images